On 18 March 2026, the UK government published its long‑awaited Copyright and Artificial Intelligence Report, alongside an economic impact assessment required under the Data (Use and Access) Act 2025. The Report sets out the government’s current thinking on how UK copyright law should interact with the development and deployment of AI, particularly the use of copyrighted works in training generative AI models. It follows several years of consultation, substantial stakeholder engagement, and increased legal scrutiny, including high‑profile litigation in the UK and abroad.
For businesses, the key message is one of caution rather than immediate change. The government has decided not to introduce reforms to UK copyright law at this stage, citing a lack of consensus between rightsholders and technology developers and insufficient evidence on the economic impact of reform. Instead, it has committed to further evidence‑gathering and monitoring of both domestic and international developments. In the short term, this means the existing legal framework remains in place, with ongoing uncertainty around risk allocation and compliance when training or deploying AI systems.
Significantly, the government has formally moved away from its earlier proposal to introduce a broad text and data mining exception with an opt‑out for rightsholders. Strong opposition from the creative industries and mixed reactions from technology companies led the government to conclude that there is currently no agreed alternative approach. Businesses relying on large‑scale data ingestion therefore cannot expect a new statutory safe harbour in the near future.
The Report does, however, signal likely reform in one area. The government proposes removing copyright protection for works generated entirely by computers without human input, on the basis that copyright should reward human creativity. Works created with the assistance of AI would continue to attract protection. While no timetable has been set, this proposal may affect business models that depend on exclusive rights in fully machine‑generated outputs.
Rather than mandatory disclosure obligations, the government intends to priorities voluntary transparency measures, technical standards, and the development of licensing solutions. It is also actively considering additional protections against non‑consensual digital replicas (deepfakes), with a further consultation expected this summer. Overall, the Report reflects a “wait and see” strategy, leaving businesses to navigate a complex and evolving risk landscape while policy continues to develop.